Subject: Re: impolitic statements (business model clues for the clueless
From: Steve Mallett <>
Date: Wed, 17 Oct 2001 03:43:27 -0300

On Wednesday 17 October 2001 13:22, Norbert Bollow wrote:
> Tom Lord <> wrote:
> >        Steve Mallett *tweaks*:
> >        > I'm curious what you are under the *delusion* you'd pay for if
> >        > it were provided from such websites as SF & FM?
> >
> > Software tool maintainers doing a coherent and mature job.
> What's the size of this market?
> What would be the price of the service?
> What marketshare can the proposed webservice business reasonably
> expect to achieve, at this price, given that there is
> competition from the do-it-yourself approach as well as
> and ?
> Based on this, will there be enough revenue to make it work out
> economically?  (I doubt it.)
> Greetings, Norbert.

Before this gets too far.... [Warning, this is also, kind of, a sales pitch]

I've given the quality of product listings a lot of thought.  And I mean... a 

Let me give you the conclusion I've come to first then explain how I got 
there: Quality Assurance of listings (at any site) is an unsexy job, and 
hence people do not do it well if at all.  It must be done by the author, but 
they would, generally, rather build something else than document what exists. 
So, it must simply be made easier to do.  That what my trovesend project is 
trying to do.  Anyway...

Factoid:  Product listings -specifically- and packaging is often out of date, 
inconsistant across sites and/or lacks more than the superficial amount of 
meta-data.   Pick any combination.  This is problematic, obviously.

How do you address this: 
1) Expect directory maintainers to examine every product listed.  That isn't 
going to happen.  Aside from the financial concerns of hiring more people to 
do this aren't you really just passing the buck to another level of 
incompetance?  That is to say, how can someone know more about a product than 
the author/maintainer?  They cannot.
2) Ask the author/maintainers to pick up the slack.  This is a good one.  Not 
unreasonable either, *but when you have an author who is prolific his product 
listings become worse with the number of prdoducts they work on and the more 
active they are in developing them. 
3) Enlist maintainers for products!  That could become a solution as more 
non-programmers want to help their favorite product's author, but how many 
maintainers do kde/gnome/licq/gimp/apache..... really need?  (emphasis on big 
and obvious products is intentional).

So, the author is the obvious choice to keep in charge of maintenance, it 
must be made easier for them to do so I'm working on building this client now.

Here's how I envision this working:
Author fills the data in (most stuff won't change like trove categories, 
author, etc), chooses where to send the data to (any directory can except the 
data if they want), the *client* goes to those sites/databases & reports 
inconsistancies back to author and what is there currently (should match up), 
author corrects and/or makes changes based on progress of the software and 
then the *client* goes out unto the opensource world, makes necessary changes 
to the product information and reports back that 'it is done, my master'.

So, it's easier to keep up to date, the blanks are filled in once for any 
site, the data is consistant regardless of where someone finds it, and you 
can likely get the author to fill out more, in-depth data about the product 
since he only has to do it once.

As an aside, I hate the _idea_ of paying someone to do a job that I expect to 
be done correctly in the first place.
Steve Mallett | Stable, Open-Source Apps |
Voice: 1.506.857.9050

Project-Listing Maintenance In A Can:  (Aug 15th/01,
I have nothing to do with license approval.)

"Non-cooperation with evil is as much a duty as cooperation with good."
        -- Mohandas Gandhi